Object detection system, object detection device, and object detection method

Abstract

An object detection system includes a first detection device configured to control a first camera to acquire a first image of a container and detect an object in the container based on the first image, and a second detection device configured to control a second camera to acquire a second image of the container. The second image is a more detailed or accurate image than the first image for detecting the object in the container based on the second image. The first detection device is further configured to train a learning model for detecting the object in the container based on the first image using detection results from the second detection device as an indication of the correct result, and then use the trained learning model for detecting the object in the container based on the first image.

What is claimed is: 
     
       1. An object detection system, comprising:
 a first detection device configured to:
 control a first camera to acquire a first image of a container, and 
 detect an object in the container based on a learning model to which the first image is supplied as input; and 
 
 a second detection device configured to:
 control a second camera to acquire a second image of the container, the second image including more data than the first image, 
 detect the object in the container based on the second image, and 
 output a detection result, wherein 
 
 the first detection device is further configured to:
 train the learning model using the detection result from the second detection device as a correct detection result. 
 
 
     
     
       2. The object detection system according to  claim 1 , further comprising:
 a sensor configured to output a signal to the first and second detection devices when the container is placed in a predetermined area. 
 
     
     
       3. The object detection system according to  claim 1 , wherein the first image is a two-dimensional image. 
     
     
       4. The object detection system according to  claim 3 , wherein the second camera is a stereoscopic camera, and the second image is a three-dimensional image. 
     
     
       5. The object detection system according to  claim 1 , wherein the first detection device is further configured to determine whether a learning achievement level of the learning model has reached a predetermined threshold level. 
     
     
       6. The object detection system according to  claim 5 , wherein the second detection device can be disconnected after the learning achievement level has reached the predetermined threshold level. 
     
     
       7. The object detection system according to  claim 5 , wherein the first detection device is further configured to, when the learning achievement level has reached the predetermined threshold level, output a detection result based on the learning model being supplied with the first image. 
     
     
       8. The object detection system according to  claim 7 , further comprising:
 a reporting device configured to report detection results from either the first detection device or the second detection device. 
 
     
     
       9. The object detection system according to  claim 8 , wherein the detection result that is output by the second detection device is reported until the learning achievement level has reached the predetermined threshold level. 
     
     
       10. The object detection system according to  claim 8 , further comprising:
 a switch configured to switch between a first state in which the detection result from the first detection device is transmitted to the reporting device and a second state in which the detection result from the second detection device is transmitted to the reporting device. 
 
     
     
       11. The object detection system according to  claim 10 , wherein the first detection device is further configured to control the switch to be in the first state after the learning achievement level has reached the predetermined threshold level. 
     
     
       12. The object detection system according to  claim 11 , wherein the first detection device is further configured to control the switch to be in the second state until the learning achievement level reaches the predetermined threshold level. 
     
     
       13. An object detection apparatus, comprising:
 a first detection device configured to detect an object in a container based on output from a first type sensor and a learning model to which the output from the first type sensor is supplied; and 
 a deep learning unit configured to supply the learning model with detection results from a second detection device configured to detect object in the container based on output from a second type sensor which is more accurate than the first type sensor, the detection results from the second detection device being consider as correct results for the learning model. 
 
     
     
       14. The object detection apparatus according to  claim 13 , wherein
 the first type sensor is a camera acquiring a two-dimensional image, and 
 the second type sensor is stereoscopic camera. 
 
     
     
       15. The object detection apparatus according to  claim 13 , further comprising:
 the second detection device. 
 
     
     
       16. The object detection apparatus according to  claim 13 , further comprising:
 a convergence determination unit configured to determine whether detections of the object by the first detection device using the output from a first type sensor and the learning model has reached a threshold level of convergence with detections of the object from the second detection device. 
 
     
     
       17. The object detection apparatus according to  claim 16 , further comprising:
 a switch configured to cause the detection result from the second detection device to be reported to an operator until the learning model has reached the threshold level of convergence and then cause the detection result from the first detection device to be reported to the operator after the learning model has reached the threshold level of convergence. 
 
     
     
       18. The object detection apparatus according to  claim 13 , further comprising:
 a reporting unit configured to notify an operator of a detection result from the first detection device or the second detection device. 
 
     
     
       19. The object detection apparatus according to  claim 18 , wherein the reporting unit comprises a lamp. 
     
     
       20. An object detection method, comprising:
 controlling a first camera of a first detection device to acquire a first image of a container; 
 controlling a second camera of a second detection device to:
 acquire a second image of the container, the second image including more data than the first image, 
 detect a left-behind object in the container based on the second image, and 
 output a detection result of the left-behind object; and 
 
 controlling the first detection device to:
 train a learning model for detecting the left-behind object in the container based on the first image; and 
 detect the left-behind object in the container based on the first image using the trained learning model.